Virtual assistant mood tracking and adaptive responses

Abstract

Among other things, embodiments of the present disclosure can help improve the functionality of virtual assistant (VA) systems by recognizing and tracking a user's mood and adapting its responses accordingly. Embodiments of the present disclosure may utilize data in real-time or near-real-time to identify a user's mood, as well as tracking a user's preferences and reactions in past interactions with the VA or in other contexts.

What is claimed is: 
     
       1. A system comprising:
 a processor; and 
 
       memory coupled to the processor and storing instructions that, when executed by the processor, cause the system to perform operations comprising:
 receiving an input from a first user directed to a virtual assistant operating on the system, the virtual assistant adapted to assist the first user in financial matters; 
 comparing the received input to a previously-received input from the first user; 
 determining an occurrence of an event associated with the first user, the event including one or more of: 
 
       a status of a financial account of the first user or a status of the first user in achieving a financial goal;
 determining social media content posted by the first user regarding the event; 
 determining a second user that is a friend or family member of the first user and is likely to have an effect on a mood of the first user; 
 determining a mood for the second user that is associated with the first user; 
 predicting the mood of the first user based on the comparison, the mood for the second user, the social media content posted by the first user regarding the event, and the event; 
 generating a plurality of responses to the received input based on the predicted mood of the first user; 
 determining, for each particular one of the plurality of responses, a probability that the response will be well received by the first user, the probability using a machine-learned model trained to find a correlation between receptivity and mood; 
 selecting a response from the plurality of response that has a highest probability that the response will be well received; and 
 providing the selected response to the first user via the virtual assistant. 
 
     
     
       2. The system of  claim 1 , wherein the input includes one or more of: a request for information from the virtual assistant, or a request for the virtual assistant to perform a task. 
     
     
       3. The system of  claim 1 , wherein the input includes one or more of:
 audio input, and text input. 
 
     
     
       4. The system of  claim 1 , wherein predicting the mood of the first user is further based on data regarding the first user received from a sensor in communication with the system. 
     
     
       5. The system of  claim 4 , wherein the sensor includes one or more of: a blood pressure sensor, a body temperature sensor, a heart rate monitor, and a sleep monitor. 
     
     
       6. The system of  claim 1 , wherein generating the plurality of responses includes identifying a predetermined time to provide the response to the first user. 
     
     
       7. The system of  claim 1 , wherein determining the mood of the first user is further based on one or more of:
 ambient noise in an environment of the first user, a room temperature in the first user's environment a social media post by the first user, and demographic information for the first user. 
 
     
     
       8. A method comprising:
 receiving, by a computer system, an input from a first user directed to a virtual assistant operating on the computer system, the virtual assistant adapted to assist the first user in financial matters; 
 comparing, by the computer system, the received input to a previously-received input from the first user; 
 determining an occurrence of an event associated with the first user, the event including one or more of: 
 
       a status of a financial account of the first user or a status of the first user in achieving a financial goal;
 determining social media content posted by the first user regarding the event; 
 determining a second user that is a friend or family member of the first user and is likely to have an effect on a mood of the first user; 
 determining a mood for the second user that is associated with the first user; 
 predicting the mood of the first user, by the computer system based on the comparison, the mood for the second user, the social media content posted by the first user regarding the event, and the event; 
 generating, by the computer system, a plurality of responses to the received input based on the predicted mood of the first user; 
 determining, for each particular one of the plurality of responses, a probability that the response will be well received by the first user, the probability using a machine-learned model trained to find a correlation between receptivity and mood; 
 selecting a response from the plurality of response that has a highest probability that the response will be well received; and 
 providing, by the computer system, the selected response to the first user via the virtual assistant. 
 
     
     
       9. The method of  claim 8 , wherein the input includes one or more of:
 a request for information from the virtual assistant, or a request for the virtual assistant to perform a task. 
 
     
     
       10. The method of  claim 8 , wherein predicting the mood of the first user is further based on data regarding the user received from a sensor in communication with the system. 
     
     
       11. The method of  claim 10 , wherein the sensor includes one or more of: a blood pressure sensor, a body temperature sensor, a heart rate monitor, and a sleep monitor. 
     
     
       12. The method of  claim 8 , wherein generating the plurality of responses includes identifying a predetermined time to provide the response to the first user. 
     
     
       13. The method of  claim 8 , wherein determining the mood of the first user is further based on one or more of:
 ambient noise in an environment of the first user, a room temperature in the first user's environment, a social media post by the first user, and demographic information for the first user. 
 
     
     
       14. A non-transitory computer-readable medium storing instructions that, when executed by a computer system, cause the computer system to perform operations comprising:
 receiving an input from a first user directed to a virtual assistant operating on the computer system, the virtual assistant adapted to assist the first user in financial matters; 
 comparing the received input to a previously-received input from the first user; determining an occurrence of an event associated with the first user, the event including one or more of: 
 
       a status of a financial account of the first user or a status of the first user in achieving a financial goal;
 determining social media content posted by the first user regarding the event; 
 determining a second user that is a friend or family member of the first user and is likely to have an effect on a mood of the first user; 
 determining a mood for the second user that is associated with the first user; 
 predicting, the mood for the first user based on the comparison, the mood for the second user, the social media content posted by the first user regarding the event, and the event; 
 generating a plurality of responses to the received input based on the predicted mood of the first user; 
 determining, for each particular one of the plurality of responses, a probability that the response will be well received by the first user, the probability using a machine-learned model trained to find a correlation between receptivity and mood; 
 selecting a response from the plurality of response that has a highest probability that the response will be well received; and 
 providing the selected response to the first user via the virtual assistant.